1. The preparation and presentation of a complete statement of financial condition, the balance-sheet, and a statement of operating results, usually called the profit-and-loss statement, are generally recognised as primary aims of accounting. Although these statements apply to the total business of an enterprise, the information which they provide may be adequate for small concerns engaged in a single line of activity. Large industrial organisations require more detailed information. The manufacture of complicated products in a multiplicity of forms and sizes created the need for the determination and analysis of the unit costs of production, and the whole science of cost accounting was developed. The development of worldwide organisations for the distribution of the products of single companies has now created a similar need for the detailed analysis of distribution costs and for the determination of operating results both by types of product and by territorial divisions. An accounting or statistical technique to meet this need will inevitably be developed, but it will not be the work of one man or of a single year. When accounting theory and practice have reached this point in their development, or when suitable statistical methods have been introduced, it will be possible to allocate business profits with some degree of assurance. Taxes, however, must be paid in the meanwhile, and it will therefore be necessary to work out equitable, if not exact, methods of allocation based on accounting procedures now in actual use.


2. In the replies to questionnaires issued at the beginning of the study of allocation methods, the preference for separate accounting for branches of foreign enterprises, not only among taxpayers, but also among tax authorities and accountants, was so marked as to create a strong presumption in its favour. The advantages of separate accounting from the point of view of the taxpayer are clear enough. The flexibility of the method makes it possible to introduce a system of accounting which meets the peculiar requirements of each branch, no matter what its size or function, and the acceptance of these branch accounts for tax purposes effectively prevents the double taxation of income, since under no system of accounting would the same item of income be allocated to two or more branches. The acceptance of branch accounts, moreover, relieves the taxpayer of the necessity of including figures for the entire enterprise on the return of each branch. To the taxpayer about the only disadvantage of separate accounting is the possibility that the keeping of separate accounts for each branch in a manner which will meet tax requirements may involve additional book-keeping expense.

3. American corporations will have a double reason for keeping accounts which show the profits earned in each foreign country separately. Section 131 (c)1 of the United States Revenue Act of 1932 now provides that, in order to obtain the credit for foreign taxes, a corporation must report, not only the total amount of profit arising from sources without the United States, but also the amount originating in each foreign country. Since branch profits must be separately determined both in computing the tax due to foreign countries and in ascertaining the credit for foreign taxes, it will be advantageous to American corporations in particular to use branch accounting methods which all countries can accept. The failure to do so may mean double taxation, or the loss of the credit for foreign taxes, or both.


4. From the point of view of the tax authorities of countries in which branches of foreign enterprises are located, returns based on a separate accounting are convenient because they relate solely to operations within the country and can be handled and verified in much the same way as the returns of independent concerns. By contrast, returns which relate to the entire business of an international enterprise are inevitably complex and not subject to the usual methods of verification. Separate accounting would be the ideal method were it not for the fact that the amount of profit reported by a branch may depend so largely on the pricing of intra-company transactions as to reake the results subject to question. But even where the amount of branch profit is significantly influenced by inter-branch charges for goods and services, some countries, notably the United States and the United Kingdom, seek to adjust these charges to a reasonable basis rather than to use any method of fractional apportionment. Experience has shown that such adjustments can usually be made on some reasonable basis. It is probable, moreover, that, if separate accounting is generally accepted, business concerns will devote more attention to the introduction of accounting methods which will reasonably reflect branch profits without the necessity of important adjustments.

5. The preference for separate accounting in principle, though marked, is not unanimous. The Spanish administration in particular advocates the method of fractional apportionment and a strong case in its favour is presented by Dr. Agust’n $$$Vinuales in his report on the Tax System and Allocation Methods in Spain.1 The validity of most of the arguments2 for the Spanish system is at once apparent if the principle of taxing branches according to the capacity of the entire enterprise to pay be accepted. Under the Spanish system, capacity to pay is measured by the total profit of an enterprise, and a share of this total profit is attributed to the branches in Spain in accordance with their estimated relative economic importance to the enterprise as a whole. A corporation, for example, which earned no profit on its total business would presumably pay no income-tax in Spain (except minimum payments on capital), even though the Spanish branches unquestionably earned substantial profits. Taxation by such methods may seem logical, but the principle of taxing foreign enterprises according to ability to pay, as interpreted by the Spanish authorities, is not generally accepted.

note note

6. In most countries, the intention is to tax all business income in the country of origin without regard to profits or losses elsewhere. This result cannot ordinarily be obtained by the method of fractional apportionment. An apportionment fraction may quite properly be used by agreement to make a reasonable division of profit among several tax jurisdictions, but it will not, except by accident, allocate profit to the place of origin. The two things are quite different. It is entirely reasonable and proper for two countries to agree that the profits of enterprises operating in both shall be divided for tax purposes in, say, the ratio of tangible property in the one to tangible property in the other. It is not reasonable to assume, however, that the amount of profit really originating in either country can be determined by this ratio, or by a ratio based on any other common factor or combination of factors. All general apportionment fractions as a matter of mathematical necessity allocate profits or losses evenly to all branches. By their very nature, they cannot allocate profits to some and losses to others. They cannot, therefore, be said to allocate profits and losses to the branches at which they originate, for at different times in almost any large enterprise some branches are profitable and some unprofitable. This fact alone should be sufficient to indicate that the method of fractional apportionment cannot meet the needs of countries which insist on taxing all income originating within the country without regard to the total profit or loss of an enterprise as a whole.

7. The confusion which commonly exists in regard to the function of a general apportionment fraction is unfortunate, and may cause a considerable amount of double taxation. If the branches of a given foreign enterprise in some one country are undeniably profitable, the tax authorities are likely to insist on levying a tax on the profits even though the company as a whole may operate at a loss. The same authorities may at the same time use an apportionment fraction to levy an income-tax on the unprofitable branches of another enterprise which happened to show a profit on its total business. This is inconsistent. Either the branch accounts, if properly kept, should be accepted whether they show a profit or a loss, or the method of apportionment should be consistently used. It should be clearly understood that a country is to tax either the profits originating within its borders or a reasonable share of total profits. It is not proper that each country should exercise the option of choosing the more profitable alternative.

8. In addition to the theoretical objection that apportionment fractions do not allocate income to the place of origin, there are practical difficulties in the use of this method comparable to those which arise in separate accounting. The original data from which an apportionment fraction must be constructed are stated in different currencies which must be converted at varying rates. What, for example, is the present investment of an American concern in a building constructed for its London branch when sterling exchange was at 4.85 dollars? The fall in the value of the pound in terms of dollars did not proportionally increase English construction costs or the value of London real estate. It did, however, effectively reduce the real value of the American company’s investment. This is only one of a number of questions which might arise if the tax authorities of, let us say, France were attempting to determine the investment of the American company in fixed property outside of France for use in an apportionment fraction. The fact that such questions do not receive detailed consideration simply indicates the extent to which apportionment fractions are based on rough approximations. The use of pay-rolls as a factor raises not only exchange difficulties, but also the problem of differences in wage levels. Ought twice as much profit to be allocated to the United States, for example, simply because American wage rates expressed in dollars happen to be twice as high as the money wages in some other country? Numerous other difficulties arise in connection with the apportionment method, but they are too familiar to tax officials to require re-statement here.

9. From the point of view of the taxpayer, the objections to fractional apportionment are two. The first is the improbability that uniform fractions will ever be used in all or even in a majority of countries. The second is that, even if uniform fractions should be used, each country would have to enquire into the worldwide business operations of international enterprises. For over twenty years, income-taxes have been levied by certain American States, and, in that time, little progress has been made toward the introduction of uniform fractions. If uniformity cannot be obtained among the different States within a single country in which language, currency and accounting methods are essentially the same, the prospects for uniformity among nations are not bright. And the fact that taxable income is not computed by the same methods in the different countries would produce inconsistent results even if uniform fractions were used. In other words, the use of a uniform fraction in all countries, even if it were possible, would not prevent the taxation of more or less than 100 per cent of the income of a business enterprise.

10. For international enterprises, however, the necessity of reporting in every country on total world business is perhaps almost as onerous as the double taxation which may result from a lack of uniformity in apportionment fractions or in methods for computing net income. The labour of preparing such returns and obtaining the information necessary to comply with the varied requirements of so many countries is, to say the least, considerable. And some of the information with respect to distant branches, although actually irrelevant in computing income in a given country, may cause unnecessary disputes, or other complications. Corporations, it is true, can usually avoid the necessity of reporting on their entire business by organising subsidiary companies, and many corporations do use this method, but it is difficult to see wherein a country benefits from forcing business to be conducted through subsidiaries rather than branches.

11. In addition to the two principal methods of allocation which have already been discussed, there is a third group consisting of the so-called empirical methods. It is characteristic of these methods to make use of information gathered from outside sources, usually from the returns of independent concerns engaged in the same or a similar line of business. Ordinarily, an average or typical rate of gross profit or of net profit to sales in the particular line of business is determined, and this rate is then applied to the sales, sometimes called the turnover or the gross receipts, of the branch in question. Other empirical methods may represent nothing but a more or less intelligent guess as to what a branch ought to earn as indicated by certain factors, such as its investment in fixed property, the size of its inventories, and the number of its employees. These empirical methods are useful as general tests of reasonableness, but they have no valid theoretical basis, and obviously they can never eliminate double taxation.

12. After several months of almost continuous study, during which the manifold difficulties of separate accounting have become increasingly clear, it is still the opinion of the writer that this method offers the best means of obtaining satisfactory allocations of profit to branches of foreign enterprises. No other method seems to have the same opportunity to relieve business concerns from taxation on more than 100 per cent of their income and at the same time to insure the reporting of a reasonable income in all countries. It is flexible enough to meet the needs of all types of business, but this very flexibility makes it impossible to prescribe anything like a uniform system. Appropriate accounting systems must be developed by the industries themselves. The adoption of separate accounting as the preferred method by international agreement would provide a strong stimulus toward the development and introduction of accounting methods which would meet tax requirements.


13. The term separate accounting does not refer to a single well-defined method, but [rather to a whole group of methods with a common aim. Indeed, separate accounting is only a point of view or an avenue of approach to the general problem of allocating business profits on a geographical basis. The accounting methods by which the allocation of profit may be accomplished must be as diverse as the conditions within the different industries are varied. Separate accounting rests on the assumption that the different branches of an enterprise are separate business units which may be treated for accounting purposes practically as independent concerns. Some enterprises are so organised that branch profits may be determined on this basis with as high a degree of accuracy as the profits of independent concerns. Other enterprises operate as organic units in which the different branches are so closely related that a true determination of the separate profits of each branch is impossible, not merely because of the large number of factors to be considered, but chiefly because the very concept of a real separate profit for each branch is contrary to the actual facts Profits must be allocated, however, even though a business enterprise is admitted to be an indivisible organic unit. Compromises between practical necessity on the one hand and pure theory on the other are not unknown in accounting. If accounting were compelled to wait upon pure theory to provide a basis for the practical quantitative answers which business activities demand, there would be no accounting. There is, for instance, no universally accepted definition of income or profit and yet accountants must ascertain period by period the amount of profit earned by each enterprise. The absence of an exact theoretical basis means simply that a criterion of reasonableness must be substituted for scientific precision in allocating profits to branches.

14. A close parallel can be drawn between separate accounting and cost accounting. Cost accounting rests essentially on the assumption that a unit cost for each product can be found and that these unit costs multiplied by the number of articles manufactured in each class must equal the total cost of operating the factory. Now unit costs constructed in this manner are hybrids consisting of the actual cost of materials and direct labour plus an average amount of overhead. They represent, therefore, neither a true differential cost (the additional cost of manufacturing one more unit) nor a true average cost, but they have been extremely useful. Recent developments in cost accounting have tended to create a clearer understanding of the significance of unit costs, and there is a growing appreciation of the fact that there is no single unit cost for any article, but rather a whole series of unit costs depending upon the purpose to be served. A similar situation exists with respect to separate accounting. Branch profits may be computed differently for different purposes and no one figure for profit would suffice for all of them. It is necessary, therefore, to select that concept of profit which is most suitable for tax purposes and then to seek reasonable methods for making the necessary computations.

15. Income-tax laws in all countries have been framed generally for the purpose of taxing whole enterprises, not parts of enterprises. When a single enterprise, however, operates in two or more countries, the parts must be taxed separately, and simple justice requires that they be taxed on the same basis as whole enterprises. In order to accomplish this result, it will be necessary to frame special laws and administrative regulations for the taxation of branches or to require simply that branch returns be made on the same basis as the returns of independent concerns. The latter procedure would place the burden of computing branch income squarely upon the accountants and the taxpayers themselves, thereby forcing them to develop and introduce systems of accounting adequate for tax purposes. The dangers of making special laws and detailed regulations are two. It is quite inconceivable that the laws and regulations of different countries should be the same, and they would necessarily be inflexible and not easily modified to meet changing conditions, even if uniformity were secured by treaty. The advantages of separate accounting are that it is flexible and readily subject to modification as methods of accounting improve or as conditions change. Moreover, even if different concerns use different accounting methods, each concern can and ordinarily should use the same method of branch accounting in the different countries in which it does business.

16. If branches are to be taxed under the laws and regulations prescribed for whole enterprises, it is clear that the aim of all methods of allocation must be to ascertain the profit which an independent concern would have earned under identical conditions. This ideal can never be attained in practice — indeed, some compromises such as the elimination of unrealised profits in inventories seem to be desirable — but there are several methods by which it may be approximated. In order to treat branches as independent business entities, two things are necessary:

17. The latter of these requirements is by far the more important. The question of pricing relates to every inter-branch transaction. Indeed, if inter-branch prices are proper, and appropriate allowances are made for risks of loss and interest on borrowed capital, a satisfactory allocation of profits can be made without placing branches formally on a capitalised basis. The establishment of a definite capital account for each branch by resolution of the board of directors would help to carry out the treatment of branches as independent concerns, but it is not absolutely necessary.1